FRONTEO's self-developed AI engine "KIBIT" 3 functions
KIBIT Functions (1): Search and Classification
- Search and classification by "connotation" of sentences and words
- High-precision recognition and retrieval of polysemous words and spelling errors
In addition to simple keyword matching, AI can detect the meaning of words and entire sentences, allowing for optimal search and comment classification of documents and articles.
This is due to the high accuracy of the algorithm's ability to find mathematical closeness, or similarity, between words using an algorithm that adheres to the distributional hypothesis*. We also combine several unique methods of analysis, such as the calculation of the contribution of words to a sentence and an integrated search that preserves the characteristics of the original sentence.
* An important concept in natural language processing, the idea that meaning is formed by the surrounding words and context rather than by the words themselves.
KIBIT Function 2: Discovery and Extraction
- Able to find evidence and predictive signs from a large volume of emails and documents
- AI can analyze text as if it were reading the context.
Our proprietary algorithms can convert the senses that attorneys and auditors use to find evidence in documents, as well as the mood of the entire text, into numerical and mathematical formulas. This allows AI to audit and review large volumes of text in place of humans.
In addition to utilizing information other than text as features and automatically optimizing the weighting of words, the algorithm also incorporates a proprietary method to suppress over-learning, thereby increasing the accuracy of discovery and extraction.
KIBIT" Function 3: Network Analysis
- Can discover chokepoints and hidden dominance in business-to-business transactions
- Able to visualize the propagation of influence and the strength of connections within a network
By analyzing open source and proprietary corporate data, complex information networks can be depicted, and the flow of goods and the degree of influence propagation can be visualized using AI to enable a speedy grasp of the overall picture.
We are developing technologies and proprietary indicators to analyze various networks, such as connections among companies in supply chains, which are indispensable for economic security, shareholding relationships, and connections among researchers/research institutions based on coauthorship relationships.
